Research on Trajectory Control Strategy of Robot Arm Based on Neural Network Compensation

Abstract

The dynamics equations of the two dimensional robot arm are established, and the uncertain variables which should be considered to control the trajectory of robot arm are analyzed. The compensating control strategy based on neural networks is presented,which considers the the nonlinearity of dynamics models and the uncertainty of parameters. The structure of the presented control strategy is simple, and the nominal value is used to fix the initial weight of neural-network, which is easy to realize in industry. The simulation research shows the effectiveness of the control strategy, and the general and adaptive ability of neural network compensators.